diff options
Diffstat (limited to 'python/fatcat_tools/harvest/pubmed.py')
| -rw-r--r-- | python/fatcat_tools/harvest/pubmed.py | 10 |
1 files changed, 8 insertions, 2 deletions
diff --git a/python/fatcat_tools/harvest/pubmed.py b/python/fatcat_tools/harvest/pubmed.py index f97702c0..98bb7f00 100644 --- a/python/fatcat_tools/harvest/pubmed.py +++ b/python/fatcat_tools/harvest/pubmed.py @@ -20,6 +20,7 @@ import sys import tempfile import time import xml.etree.ElementTree as ET +import zlib from urllib.parse import urlparse import dateparser @@ -114,8 +115,13 @@ class PubmedFTPWorker: url = "ftp://{}{}".format(self.host, path) filename = ftpretr(url) with tempfile.NamedTemporaryFile(prefix='fatcat-ftp-tmp-', delete=False) as decomp: - gzf = gzip.open(filename) - shutil.copyfileobj(gzf, decomp) + try: + gzf = gzip.open(filename) + shutil.copyfileobj(gzf, decomp) + except zlib.error as exc: + print('[skip] retrieving {} failed with {} (maybe empty, missing or broken gzip)'.format( + url, exc), file=sys.stderr) + continue # Here, blob is the unparsed XML; we peek into it to use PMID as # message key. We need streaming, since some updates would consume |